Characterization of Control in a Superconducting Qutrit Using Randomized Benchmarking

TL;DR

This paper demonstrates high-fidelity control of a superconducting qutrit using randomized benchmarking, quantum process tomography, and a novel gate implementation method, highlighting decoherence as the main error source.

Contribution

It introduces a method for implementing qutrit gates with only two-state rotations and characterizes their fidelity in a superconducting circuit.

Findings

Average fidelity of 98.89% for qutrit Clifford gates

Decoherence identified as primary error source

Method applicable to any unitary operation on a qutrit

Abstract

We characterize control of a qutrit implemented in the lowest three energy levels of a capacitively-shunted flux-biased superconducting circuit. Randomized benchmarking over the qutrit Clifford group yields an average fidelity of 98.89 $\pm$ 0.05%. For a selected subset of the Clifford group, we perform quantum process tomography and observe the behaviour of repeated gate sequences. Each qutrit gate is generated using only two-state rotations via a method applicable to any unitary. We find that errors are due to decoherence primarily and have a significant contribution from level shifts. This work demonstrates high-fidelity qutrit control and outlines avenues for future work on optimal control of superconducting qudits.